Hand paralysis is one of the most common complications in stroke patients, which severely impacts their daily lives. This article presents a wearable hand rehabilitation system that supports both mirror therapy and task-oriented therapy. A pair of gloves, i.e., a sensory glove and a motor glove, was designed and fabricated with a soft, flexible material, providing greater comfort and safety than conventional rigid rehabilitation devices. The sensory glove worn on the nonaffected hand, which contains the force and flex sensors, is used to measure the gripping force and bending angle of each finger joint for motion detection. The motor glove, driven by micromotors, provides the affected hand with assisted driving force to perform training tasks. Machine learning is employed to recognize the gestures from the sensory glove and to facilitate the rehabilitation tasks for the affected hand. The proposed system offers 16 kinds of finger gestures with an accuracy of 93.32%, allowing patients to conduct mirror therapy using fine-grained gestures for training a single finger and multiple fingers in coordination. A more sophisticated task-oriented rehabilitation with mirror therapy is also presented, which offers six types of training tasks with an average accuracy of 89.4% in real time.

中文翻译：

---

带有软手套的可穿戴手部康复系统


手麻痹是中风患者最常见的并发症之一，严重影响他们的日常生活。本文介绍了一种可穿戴手部康复系统，支持镜像疗法和任务导向疗法。一副手套，即感觉手套和运动手套，采用柔软、有弹性的材料设计和制造，比传统的刚性康复装置提供更大的舒适性和安全性。戴在未受影响的手上的传感手套包含力和弯曲传感器，用于测量每个手指关节的握力和弯曲角度以进行运动检测。电机手套由微电机驱动，为受影响的手提供辅助驱动力以执行训练任务。采用机器学习来识别感官手套的手势，并促进受影响手部的康复任务。该系统提供16种手指手势，准确率高达93.32%，允许患者使用细粒度手势进行镜像治疗，训练单指和多指的协调能力。还提出了更复杂的以任务为导向的镜像疗法康复，它提供六种类型的训练任务，平均实时准确率为89.4%。